Constant energy calibration for permittivity based moisture probes

- Errors with permittivity probes are classified as primary, secondary and tertiary.
- A unique approach in considering natural variations of dry density was proposed.
- Accuracy significantly improved for ThetaProbe and WET sensor with proposed approach.

SummarySoil moisture or soil water content (SWC) is an important variable and its determination with an acceptable accuracy is an essential need for a variety of hydrological disciplines. Among the alternatives, permittivity based probes are gaining popularity and becoming established techniques. However, probes are affected by local soil conditions, requiring soil-specific calibrations for accurate and reliable measurements. Measurement errors associated with these probes can be classified in three groups as primary, secondary and tertiary. Primary errors are mostly associated with the underlying operating principles of a probe. Soil density variation is inherently the major source of secondary error as it can create a measurement error as much as 0.030 m3 m−3. The error associated with the secondary variables can be reduced by up to 0.015 m3 m−3 by targeting the half range of the density variations. Currently available techniques do not explicitly quantify/minimize the secondary errors. Tertiary errors arise mostly from user dependent factors associated with probe-soil contact discontinuities and unexpected small scale environmental variations in the vicinity of the measurement point.This study presents a unique approach in considering the natural variations of the dry density by proposing to conduct the soil specific calibrations on repacked samples under constant compaction/compression energy of 350 kN m/m3. A simple probe insertion procedure was also employed by using guides and predrilling for probe rods in order to minimize the tertiary variables. A comparative assessment of the proposed calibration approach was conducted, considering a time domain reflectometry (TDR) probe and two relatively affordable moisture probes with different operating principles (ThetaProbe ML2x and Wet-2 sensor). The findings of this study showed that accuracy levels of 0.012 m3 m−3, 0.015 m3 m−3 and 0.016 m3 m−3, through Root Mean Square Error (RMSE), were achieved for TDR, ThetaProbe and WET sensor, respectively, for a wide range of soil types including the clays and silts. It is concluded that the proposed constant energy calibration approach help contain the secondary effects due to density variation in a tight range and that the recommended probe installation procedure help minimize the tertiary effects associated with the probe-soil contact discontinuities.